[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#699913: linux-image-3.2.0-4-amd64: Memory hotplug (VMware) often fails



Am 14.02.2013 05:39, schrieb Ben Hutchings:

Hello Ben,

Squeeze is another story, but I think there is another problem as well.
Previously we sometimes saw the backtrace, sometimes just the following
message.


[   44.220000] VMCIUtil: Updating context id from 0x7a3c21d6 to
0x7a3c21d6 on event 0.
[   44.252000] Hotplug Mem Device
[   44.252000] System RAM resource 20000000 - 27ffffff cannot be added
[   44.252000] ACPI:memory_hp:add_memory failed
[   44.252000] ACPI:memory_hp:Error in acpi_memory_enable_device
[   44.252000] acpi_memhotplug: probe of PNP0C80:00 failed with error -22

This is 'invalid argument' - wonder where that's coming from?

[   44.252000]  driver data not found
[   44.252000] ACPI:memory_hp:Cannot find driver data
[   44.268000] Hotplug Mem Device
[   44.268000] init_memory_mapping: 0000000028000000-0000000030000000
[   44.268000]  0028000000 - 0030000000 page 2M
[   44.280000]  [ffffea00008c0000-ffffea0000abffff] PMD ->
[ffff88001f200000-ffff88001f3fffff] on node 0
[   44.280000] Hotplug Mem Device
[   44.284000] init_memory_mapping: 0000000030000000-0000000038000000
[   44.284000]  0030000000 - 0038000000 page 2M
[   44.284000]  [ffffea0000a00000-ffffea0000bfffff] PMD ->
[ffff88001e400000-ffff88001e5fffff] on node 0
[   44.340000] Hotplug Mem Device
[   44.340000] init_memory_mapping: 0000000038000000-0000000040000000
[   44.340000]  0038000000 - 0040000000 page 2M

We did not observe the backtrace anymore, but the "driver data not
found" is still there.

So I think the patch fixes the backtrace (allocation error) on both
squeeze and wheezy, but squeeze has a second issue. I'll go through the
bug reports and open a new one.

Right, this is definitely a separate bug.

I finally found an article that describes this behaviour. Turns out that when this message appears, hotplug does not fail completely, but one memory bank (128MB) is missing.

http://communities.vmware.com/message/1299666#1299666

I have opened Bug #700544 for this, it only seems to affect Squeeze. Don't know whether it is worth fixing that, the patch is more than a few lines and Squeeze will hopefully be oldstable soon.

Bernhard

Attachment: smime.p7s
Description: S/MIME Kryptografische Unterschrift


Reply to: